Circuit Check, lm2903's blowing

Status
Not open for further replies.

Ambient

New Member
TailLights.JPG
**broken link removed**

I have this circuit assembled and the lm2903's keep blowing. I was wondering if anyone could take a quick look and let me know if there are any problems with the design.

Some parts may seem unnecessary, and that's because the I made some changes after the PCB was designed, so some parts were retained.

I was thinking the way the transistors were wired to the comparators the was the problem. The idea behind the design is that the 5V regulator provides a stable reference and when the blinker goes on, the brake and running lights turn off.
 
Last edited:
I think a voltage regulator into a high impedance input is a bad idea. I think connecting a high impedance input to an external circuit with no current limiting is a bad idea.

I'm guessing that the open collector output is working just fine and not overly stressed at all. Is this correct or am I missing something you forgot to mention?
 
The sink current ranges from 6mA to 16mA capable according to the spec sheets. The 22k resistors are limiting the sink current to 0.66mA max when the output of the comparators is low (with supply voltage of 14.5V max). So that should not be the problem.

Oh and I am using 2n2222 transistors. I used other ones because Eagle did not have to-92 package 2n2222's.

Could the problem be that I do not have input resistances? The spec sheets mention that but do not show it in their example circuits.

I am guessing since I cannot find source current info, and since it uses pull-up resistors, that the comparators cannot supply any current, so should I just use the 22k resistor and get rid of the resistors that are on the base of the 2n2222's? This way the 22k would be the base resistor.
 
Last edited:
OK I just checked the circuit without the comparator plugged in. When I force the connection where the output pins are to GND the lights shut off as they should. So the transistors are fine. They may not be saturating fully though.
 
I did some more experimenting, it seems that the output of the comparator has problems reaching Vcc when the output has a load resistance less than the pullup resistor, which makes sense since it is just a voltage divider at that point. So adding a op-amp to buffer the voltage may help with that.

I have a feeling he comparators were blowing because I was not using input resistors. I figured they were not needed since the input impedance was high already. I breadboarded the circuit with input resistors and it performed exactly as designed.
 
I can't see where the power supply to the 2903 comes from - you haven't shown it I guess.
If it is direct from the bike or vehicle supply, it is likely to have a lot of spikes and surges on it which may damage the ICs.
 
Eagle displays power rails for IC's differently, its to the left of the lm7805. I have some filter capacitors that should be adequate to protect against noise, but I can always add more. I need the IC to be powered from the bike so that I can get the full voltage swing to the transistors. I could have used a regulator for power and a opto-isolator, but I already had the PCB's made.

If someone can tell me whether the IC was destroyed due to no input resistors or not that would be very helpful. I don't want to remove the input R on the breadboarded version and risk popping another, I am running low. Thats all I can think of, but I have used lm339's without input resistors directly to op-amps and they work fine.

I am checking the PCB now.
 
Last edited:

Using an op-amp to buffer what is supposed to be a DIGITAL output seems like a poor choice.

Wouldn't it make more sense to invert your logic, and have the comparator output sinking the base of a PNP transistor to pull the base of the NPN LED switching transistor high? Even simpler would be to move your current base resistors to where the pull-up resistor is instead. The whole point of base resistors is to limit the base current, which is already handled by the fact that you're using a pull-up resistor with an open-collector output.
 
yea I can get rid on R4, R11, R12, R5 since the output is open when its in a high state. But I am also trying to avoid changing the wiring too much since it is all on the PCB's already. I may have to just toss the PCB's I ordered and make my own double sided PCB's. It kills me to think about doing that though, they cost me $30 total for 2.

If I was to invert the logic from the comparators I would have to unsolder the IC socket and solder wires to swap the + and - inputs. I am not looking forward to doing that but if I have to I will.

EDIT:edited some mispellings
 
Last edited:
Replace the 2903 with a pair of NPN transistors and four resistors. Eliminate 7805.

Emitters to ground.
Each base in series with resistor (10K) instead of each (-) input.
Resistors (4.7K) from each base to emitters.
Collectors are outputs.
Change R9 and R10 to 4.7K.

All of this might fit into the IC socket.
 
Last edited:
hmm. Yea I over-engineered with the comparator. But hey, thats how we all learn. Thanks for the tips. Now I just gotta figure out what to disconnect on the PCB, the wiring is very unorganized due to it being my first PCB and also the light pattern of the LED's being circular.

EDIT: Does anyone have experience with what kind of noise to expect on the 12V line? It's a Honda Goldwing, early 90's I believe. I can always hook up my O-scope and check to make sure the filter caps are working properly, but knowing what to do beforehand will help. The bike voltage will, depending on engine RPM, range from 12-14.5V. Will the battery eliminate noise from the alternator? I imagine it would act like a cap and eliminate noise up to some frequency.
 
Last edited:
The battery is a pretty good filter. If this circuit is on the IGN or ACC circuit, it's disconnected at the worst possible times, during starting and at shutoff.
 
Power is directly from the original brake, running, and blinker lines. So it will not be on the ACC. And if the blinkers are wired for the emergency flashers then they will have power with the bike off.

This will be replacing some power hungry incandescents. There is a radio and GPS and other devices running on the bike, so I am trimming the fat by replacing all the bulbs with LEDs. Except for the headlight, but that may be soon when I find high enough power LEDs or a drop-in replacement.
 
Status
Not open for further replies.

Similar threads

A
  • Question
1980's electronic typewriter transformer bad?
2 3
Replies
53
Views
6K
Nigel Goodwin
  • Question
IR circuit issues
Replies
16
Views
1K
eTech
E
C
  • Locked
  • Question
Battery Charger Circuit
Replies
10
Views
2K
rjenkinsgb
J
  • Question
Variable frequency 555 circuit
2
Replies
29
Views
2K
Papabravo
M
  • Question
Kinergetic Subwoofer Circuit Analysis
2
Replies
35
Views
5K
MLVW
M
Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…